KR20170084357A - Method for decreasing immunogenicity - Google Patents

Abstract

The present invention relates to a method for reducing the immunogenicity of an antibody variable domain.

Description

Methods of reducing immunogenicity {METHOD FOR DECREASING IMMUNOGENICITY}

Related application

This application is filed on December 23, 2009 in the United States Provisional Patent Application No. 61/289,446, the United States Patent Law 35 U.S.C. Priority claims were made under §119, the content of which is incorporated herein for reference.

Field of invention

The present invention relates to methods of denaturing the immunogenicity of antibody variable domains, in particular scFvs.

A therapeutic antibody administered to a subject in need is usually recognized as a foreign object by the subject's immune system. Even when administered antibodies are humanized, e.g., by grafting mouse CDRs to the human immunoglobulin framework to minimize mouse components, they can still elicit an immune response that impairs the efficacy and/or stability of the therapeutic agent. .

According to the literature, the patient's antibody response depends on the presence of B-cell epitopes and T-cell epitopes. When the B-cell receptor recognizes and binds to an antigen such as an administered therapeutic antibody, the antigen is internalized to the B-cell by receptor-mediated endocytosis to perform a proteolytic process. The peptides then obtained are provided by MHC class II molecules. When a T helper cell recognizes a T cell epitope, the T helper cell stimulates the corresponding B cell to proliferate and differentiate into an antibody-producing plasma cell.

In order to reduce the response of the patient's immune system to the administered antibody, various de-immunization techniques have been provided in the prior art. While most of the current technologies focus on T-cell removal, there are only limited examples of how to reduce B-cell immunogenicity.

International Publication No. WO 93/18792 describes a method of denaturing an antibody by partial reduction of the antibody. While this selectively weakens their ability to denature their immunogenicity to elicit anti-isotypic responses, they can still induce anti-isotypic responses. Although this method is suitable for vaccines, anti-isotype responses are not desirable for other therapeutic applications.

Molineux G (2003) Pharmacotherapy 23: 35-85 describes the coupling of proteins to high molecular weight polyethylene glycols. However, Onda, M. et al. [PNAS Vol 105(32): 11311-11316 (2008)] reported limited success of this method using hybrid proteins consisting of variable fragments bound to bacterial toxins or plant toxins. Their hydride proteins were inactivated; In addition, they only confirmed a slight decrease in immunogenicity.

The second method consists of chemotherapy prior to antibody administration, where the patient is treated with cyclophosphamide or fludarabine. These treatments are not desirable for patients because they damage the immune system (Kusher, BH et al (2007), Pediatr Blood Cancer 48: 430-434; Leonard JP et al (2005), J Clin Oncol 23: 5696-5704. ).

Nataga, S. and Pastan, I. (2009), Adv Drug Deliv Rev, p. In 977-985 and Onda, M. et al (2008), PNAS Vol 105(32): 11311-11316, B-cell epitope removal by point mutations in "antigenic hot spots" on the foreign protein surface. Proposed. They replaced large hydrophilic residues with a wide exposure range with small amino acids (alanine, glycine and serine). Alanine is preferred for substitution because it is typically present in hidden and exposed positions of all secondary structures, and also does not provide new hydrogen bonds. Alanine retains its antigenic form as there are no side chain atoms after the β-carbon that can react with the antibody. However, the "hot spots" described by Nataga and Pastan are conformational epitopes located in discrete clusters on the protein surface. A huge amount of experimentation is required to determine the location of an epitope that cannot be reproduced by computer simulation, and therefore their method is not a general way to reduce the immunogenicity of antibodies that can be applied conventionally. Moreover, the principle hypothesis of this method is that mainly hydrophilic residues on the surface of the molecule are involved in contact with the host antibody. For most foreign proteins this is actually true when only a portion of the naturally occurring protein (e.g. fragments, domains) is used, but contact with other domains exposes previously blocked hydrophobic amino acids to solvents, resulting in epitopes on the immune system. May be provided as This is clearly the case for Fv antibody fragments, in which case the interface residues on the variable domain are covered in the Fab fragment but exposed in the isolated variable domain. Currently available algorithms for predicting B-cell epitopes are difficult to prove and typically have a low success rate.

Therefore, there is a need in the art for simple methods to effectively reduce immunogenicity for antibody fragments, particularly variable domains.

Summary of the invention

Therefore, it is a general object of the present invention to provide a method for reducing the immunogenicity of antibody variable domains without performing a large amount of molecular modeling work. In particular, it is an object of the present invention to provide a method for removing B-cell epitopes from antibody variable domains.

Accordingly, the present invention includes the step of substituting one or more amino acid residues of a variable light chain and/or a variable heavy chain, wherein the residues are the corresponding full-length antibody or variable chain of Fab and a constant interchain boundary It provides a method of reducing the immunogenicity of an antibody variable domain comprising a variable light chain and/or a variable heavy chain present in

In one aspect, the antibody variable domain is an scFv, Fv fragment or single domain antibody, in particular an scFv.

In one aspect, one or more amino acid residues of the variable light chain and/or variable heavy chain to be substituted are consensus residues of each subtype.

In another aspect, the one or more amino acid residues to be substituted are leucine (L), valine (V), aspartic acid (D), phenylalanine (F), arginine (R) and/or glutamic acid (E).

In any aspect, the one or more amino acid residues of the variable light chain are at positions 99, 101 and/or 148 (AHo numbering). In another aspect, the one or more amino acid residues of the variable heavy chain are at one or more of the 12, 97, 98, 99, 103, and/or 144 positions (AHo numbering).

In another aspect, the one or more amino acid residues to be substituted in the variable heavy chain include (a) leucine (L) at amino acid position 12 of the heavy chain; (b) valine at position 103 of the heavy chain amino acid (V); And/or (c) leucine (L) at position 144 of the heavy chain amino acid.

In another aspect, the present invention provides an antibody variable domain obtainable by the method described herein, and a pharmaceutical composition comprising the antibody variable domain.

According to the detailed description below, the present invention may be better understood and objects other than the above may become apparent. This description is made with reference to the accompanying drawings.
1A shows a schematic diagram of a bridge ELISA used to detect a preexisting anti-scFv antibody. 1: plate surface, 2: scFv903; 3: Anti-drug antibody (ADA); 4: biotinylated scFv903; 5: Streptavidin Poly-HRP (Hose Radish Peroxidase).
1B shows the principle of confirmation evaluation in which ADA bound to a drug and biotinylated scFv903 competes with excess scFv903, 34 max (791), scFv903_DHP (961) and scFvl05 (100 mcg/ml).
Figure 2 shows the signal intensity of 149 individual serums in a colorimetric assay (Bridge ELISA) for detecting anti-scFv 903 antibodies. Signals above the assay cut point indicate the presence of anti-scFv 903 antibody (ADA) in each serum sample. Approximately 30% of the test serum samples were positive in the assay.
3 shows the variable light chain sequence arrangement of the solvent exposure sites of four different scFvs. Upper panel: each amino acid in scFv 903 and the amino acids of each scFv of different types are shown in bold. The lower panel shows the epitope category with which each individual location is associated and the percentage of human serum showing binding to each epitope category.
4 shows the variable heavy chain sequence arrangement of the solvent exposure sites of four different scFvs. Upper panel: each amino acid in scFv 903 and the amino acids of each scFv of different types are shown in bold. The lower panel shows the epitope category with which each individual location is associated and the percentage of human serum showing binding to each epitope category.
5 shows a molecular structure model of scFv 903. 5a: front view; 5b: 180° rotation. Gray: residues potentially participating in epitope category β; Black: residues potentially participating in epitope category α.

In order to more easily describe the present invention, specific terms were first defined. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs. Although methods and materials similar or equivalent to those described herein may be used in the practice or testing of the present invention, suitable methods and materials are described below. All publications, patent applications, patents and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. Further, the materials, methods, and examples are for illustrative purposes and are not meant to be limiting.

The expression "immunogenicity" as used herein refers to the occurrence of an antibody epitope on a B-cell or a protein administered to a subject, whereas such B-cells or antibodies (also called anti-drug antibodies: ADA) are It may be present prior to administration of the protein.

This range of immunogenicity can be determined by ELISA assays and expressed as a percentage of human serum comprising a measurable amount of pre-existing ADA. The decrease in immunogenicity between the protein and the corresponding protein engineered for the purpose of reducing its immunogenicity was compared to the percentage of serum samples containing ADA for the engineered protein compared to the percentage of serum samples containing ADA for the native protein. Can be measured. A lower number or percentage of positive serum samples for the engineered protein indicates a decrease in immunogenicity for the engineered protein. Competitive ELISA setups are used for more sensitive measurements that can be applied based on a single serum sample. In this competitive ELISA, the engineered protein competes with the native protein for binding of ADA in the test serum. The lower the ability of the engineered protein to compete with the native protein, the more successful the immunogenicity decline.

Preferably, the range of immunogenicity reduction is referred to as the percentage of serum samples in which the engineered protein is no longer able to effectively compete with the native protein. Effective competition is defined as a threshold (relative signal in competitive ELISA), -100 represents a perfect competitor (no reduction in immunogenicity) and 0 represents no competition (complete absence of ADA epitopes). Typically these thresholds for effective competition may be -90, -80, -70, -60, -50, -40, -30, -20, -10, or> -10.

As used herein, “interface” or “interface-interface” refers to regions located between the variable domain and constant region 1 (CL1 or CH1) of a full-length antibody or between the Fab portion and the Fc domain (CH2 and CH3).

As used herein, "ADA" is an abbreviation of anti-drug antibody, which refers to a pre-existing antibody in the serum or serum of a patient.

The term "antibody variable domain" (V-Domain) refers to a molecule containing all or part of the antigen binding site of an antibody, for example all or part of the heavy and/or light chain variable domains, and thus the antibody variable domain is Specifically, it recognizes the target antigen. Therefore, this term corresponds to the V-J-REGION or V-D-J-REGION of an immunoglobulin. This V-Domain was designated as VL (V-Domain of Ig-light chain) or VH (V-Domain of Ig-heavy chain). Non-limiting examples of antibody variable domains are as follows:

(i) an Fv fragment comprising _{the V L} and V _{H domains of an antibody single arm,}

(ii) single chain Fv fragment (scFv),

(iii) Dab fragments consisting of VH or VL domains (Ward et al., (1989) Nature 341:544-546), Camelid (Hamers-Casterman, et al., Nature 363:446-448 ( 1993), and Dumoulin, et al., Protein Science 11: 500-515 (2002)) or a single domain antibody such as a Shark antibody (eg, Shark Ig-NARs Nano-bodies®).

The term “antibody framework” or “framework” as used herein refers to a portion of a variable domain, VL or VH that acts as a scaffold for the antigen binding loop of the variable domain (Kabat, EA et al. ., (1991) Sequences of proteins of immunological interest.NIH Publication 91-3242).

The term "antibody CDR" or "CDR" as used herein is described in Kabat E.A. et al., (1991) Sequences of proteins of immunological interest. NIH Publication 91-3242] refers to the complementarity determining region of an antibody consisting of an antigen-binding loop. Each of the two variable domains of the antibody Fv fragment is, for example, three CDRs.

The term "single chain antibody" or "scFv" refers to a molecule comprising an _{antibody heavy chain variable region (V H} ) and an antibody light chain variable region (V _{L) linked by a linker.} Such scFv molecules may have the formula: NH ₂ -V _L -linker-V _H -COOH or NH ₂ -V _H -linker-V _L -COOH.

The term "subtype" refers to a set of V-DOMAINs that belong to the same group and share a high percentage of identity within a given species. The term “subtype” refers to a subtype defined by each consensus sequence as defined in Knappik (2000). "Subfamily" or "subclass" is used synonymously with "subtype". The term "subtype" as used herein refers to a sequence that shares a very high degree of identity and similarity with each consensus sequence representing their subtype. For the "subtype" to which a particular variable domain belongs, all are determined by aligning each sequence with a known human germline segment or a defined consensus sequence of each subtype and subsequent binding to a particular subtype based on the greatest homology. . Methods for determining homology and grouping sequences using a search matrix such as BLOSUM (Henikoff 1992) are known to those of skill in the art.

The “consensus residue” at a given position can be determined by producing an amino acid consensus sequence of a given subtype. The term "amino acid consensus sequence" as used herein uses a matrix of at least two, preferably more, aligned amino acid sequences and allows gaps in the arrangement to be generated so that the most frequent amino acid residues can be determined at each position. Refers to the amino acid sequence. The consensus sequence is a sequence that contains the amino acids that appear most frequently at each position. When two or more amino acids appear equally at a single position, the consensus sequence includes both or both of these amino acids. The amino acid sequence of a protein can be analyzed at various levels. For example, conservancy or variability can be exhibited at a single residue level, multiple residue levels, multiple residues with gaps, and the like. Residues can be conservative of the same residue or can be conserved at the class level. Other classes are known to those of skill in the art and can be defined using other data to assess structure determination or substitutability. In this case, an amino acid that can be substituted may refer to an amino acid that can be substituted at that position and maintain functional preservation. As used herein, when aligning one amino acid sequence (e.g., a first VH or VL sequence) to one or more additional amino acid sequences (e.g., one or more VH or VL sequences in a database), one An amino acid position in a sequence (eg, a first VH or VL sequence) can be compared to a “corresponding position” in one or more additional amino acid sequences. As used herein, “corresponding position” refers to an equivalent position within a sequence that is compared when the sequences are optimally aligned, ie when the sequences are aligned to obtain the highest percentage of identity or similarity.

The AHo numbering system used throughout the detailed description is described in A. Honegger and A. Pluckthun (2001), J.Mol.Biol. 309: 657-670.

The term "patient" refers to a human or non-human animal.

The terms “treat”, “treatment” or “treatment” are intended to alleviate or prevent, cure, delay, or alleviate the severity of one or more symptoms of a disease or recurring disease, or are expected to be absent such treatment. It refers to a method of treatment and/or prevention to prolong the survival rate of a subject more than that.

“Hydrophilic” amino acids are polar and electrically charged amino acids such as Asp, Glu, Lys, Arg and His.

Polar and uncharged amino acids are Gly, Ser, Thr, Cys, Asp, Gin and Tyr.

"Hydrophobic" amino acids are typically non-polar amino acids, such as Ala, Val, Leu, Ile, Met, Phe, Trp and Pro.

In a first aspect, a method of reducing the immunogenicity of an antibody variable domain is described. The antibody variable domain comprises a variable light chain and/or a variable heavy chain, the method comprising substituting one or more amino acid residues of the variable light chain and/or variable heavy chain, wherein the residues are the corresponding full-length antibody (or Fab, i.e. It exists at the interface between the variable chain and the constant chain of an antibody or an antibody fragment comprising a constant domain or a portion thereof).

One or more amino acid residues selected for substitution are preferably present at the interface between the variable and constant chains of the corresponding full-length antibody (or Fab, i.e. antibody or antibody fragment comprising a constant domain or part thereof), and an antibody such as scFv They are solvent exposed in the variable domain. The interface is also called the V/C domain interface.

Antibody variable domains are, for example, scFvs, Fv fragments or single domain antibodies, preferably scFvs.

Of particular note are the amino acid residues at positions that form discontinuous, i.e., morphological B-cell epitopes. These residues are those found at the following positions (AHo numbering):

Variable light chain positions 99, 101 and/or 148; And

Variable heavy chain position 12, 97, 98, 99, 103, and/or 144.

The 12, 98, 103, and 144 residue positions of the heavy chain (AHo numbering) as well as the 99, 101 and 148 residue positions of the light chain (AHo numbering) are described in Nieba et al. (1997) Protein Eng., Apr;10(4):435-44 (also described in US Pat. No. 6,815,540). Nieba proposed the substitution of a hydrophilic amino acid with a hydrophilic amino acid at the indicated position, but this document does not address the effect of such substitution on the immunogenicity of the molecule. In addition, the authors emphasize that not all of these hydrophobic residues are equally good candidates for replacement. The presence of hydrophobic patches is preserved in all antibodies, but their exact location and extent are different.

As is known in the art, in particular

(i) exist in the turn area of the secondary structure,

(ii) having an elastic large side chain or a bulky side chain, or

(iii) Hydrophobic amino acids can easily become part of a B-cell epitope, inducing an immunogenic response. Elimination of the immunogenic amino acids can interfere with the B-cell epitope, thereby enhancing the tolerance of the patient's antibody variable domains.

Preferably, the one or more amino acid residues selected are substituted with an amino acid that is less immunogenic than the selected amino acid, i. These low immunogenic amino acids are those with reduced ADA reactivity compared to ADA reactivity against antibody variable domains containing native (ie, unsubstituted) amino acids.

Immunogenicity, i.e., the property of inducing an antibody response in the body of a patient, can be estimated, for example, by its antigenicity, i. Antigenicity can be measured using serum from a donor containing potentially pre-existing antibodies, for example by bridge ELISA (see Examples and Figure 1) with ADA reactivity. Therefore, for evaluation of lower immunogenic amino acids, antibody variable domains can be mutated at the indicated positions. This mutant effect on immunogenicity can be assessed by competing the signal of the progenitor antibody in a bridge ELISA with its engineered derivatives, perhaps less immunogenic, as described herein. Binding of ADA to an antibody can also be assessed using unlabeled binding assays such as surface plasmon resonance, fluorescence resonance energy transfer (FRET), colorimetric assays, and the like.

In one embodiment, the amino acid selected for substitution is at one or more positions selected from the group consisting of variable light chain residues 99, 101 and 148 and variable heavy chain residues 12, 97, 98, 99, 103 and 144.

Preferred amino acids selected for substitution are surface exposed but hidden by the corresponding full-length antibody or constant domain within the Fab.

In one embodiment, one or more amino acid residues of the variable light chain and/or variable heavy chain to be substituted are consensus residues of each subtype. For example, preferred amino acids selected for substitution are leucine (L), valine (V), aspartic acid (D), phenylalanine (F), arginine (R), and/or glutamic acid (E).

More preferably, the one or more amino acid residues selected for substitution are selected from the group consisting of variable light chain residues D99, F101 and L148 and variable heavy chain residues L12, R97, A98, E99, V103 and L144.

Even more preferably, leucine (L), valine (V), phenylalanine (F) and/or alanine (A) are substituted with polar amino acids, preferably serine (S) and/or threonine (T).

In particular, the DHP motif described in the international patent application PCT/CH2009/00022 unexpectedly reduces the immunogenicity of the antibody variable domain without adverse effects on the thermal stability, refolding, expression yield, aggregation and/or binding activity of the antibody variable domain. Was found. DHP motifs include amino acid residues of variable heavy chains 12, 103 and 144 (AHo numbering) in which the following amino acids are present:

(a) serine at position 12 of the heavy chain amino acid (S);

(b) serine (S) or threonine (T) at position 103 of the heavy chain amino acid; And/or

(c) Serine (S) or threonine (T) at position 144 of the heavy chain amino acid.

International patent application PCT/CH2009/00022 does not provide any clue that the disclosed denaturation is suitable for reducing the immunogenicity of antibody variable domains.

The DHP motif is located at the V/C interface of the Fab fragment and is solvent exposed upon removal of the constant domain. Thus, in a preferred embodiment of the present invention one or more amino acid residues are selected for substitution from the group consisting of variable heavy chains 12, 103 and 144 (AHo numbering). Preferably,

(a) leucine (L) is at position 12 of the heavy chain amino acid;

(b) valine (V) is at position 103 of the heavy chain amino acid; And/or

(c) Leucine (L) is present at position 144 of the heavy chain amino acid.

These residues are highly conserved in the human framework. Thus, substituting one or more of these residues provides a general way to deimmunize antibody variable domains without affecting the biophysical properties of the molecule, and the methods described herein apply to any framework of antibody variable domains. It is possible. Residues preferably present at the indicated positions are substituted by the following residues:

(a) serine at position 12 of the heavy chain amino acid (S);

(b) serine (S) or threonine (T) at position 103 of the heavy chain amino acid; And/or

(c) Serine (S) or threonine (T) at position 144 of the heavy chain amino acid.

Even more preferably it is substituted as follows: L12S, V103T and/or L144T.

The antibody variable domain can be for any target and specifically binds to the target. Examples of exemplary targets include, but are not limited to: transmembrane molecules, receptors, ligands, growth factors, growth hormones, coagulation factors, anticoagulants, plasminogen activators, serum albumin, hormones or growth factors. Receptor, neurotrophic factor, nerve growth factor, fibroblast growth factor, transforming growth factor (TGF), CD protein, interferon, colony stimulating factor (CSF), interleukin (IL), T-cell receptor, surface membrane protein, virus Protein, tumor-associated antigen, integrin or interleukin, VEGF; Renin; Human growth hormone; Bovine growth hormone; Growth hormone releasing factor; Parathyroid hormone; Thyroid stimulating hormone; Lipoprotein; Alpha-1-antitrypsin; Insulin A-chain; Insulin B-chain; Proinsulin; Follicle-stimulating hormone; Calcitonin; Luteinizing hormone; Glucagon; Coagulation factor VIIIC; Coagulation factor IX; Tissue factor (TP); von Willebrands factor; Protein C; Atrial natriuretic factor; Lung surfactant; Urokinase; Human urine; Tissue-type plasminogen activator (t-PA); Bombesin; Thrombin; Hematopoietic growth factor; Tumor necrosis factor-alpha or -beta; Enkephalinase; RANTES (Regulated on Activation Normally T-cell Expressed and Secreted); Human macrophage inflammatory protein (MIP-1)-alpha; Human serum albumin; Muellerian-inhibitors; Relaxin A-chain; Lilaxin B-chain; Prorylaxin; Mouse gonadotropin-related peptide; Microbial protein, beta-lactamase; DNase; IgE; Cytotoxic T-lymphocyte-related antigen (CTLA); CTLA-4; Inhibin; Activin; Vascular endothelial growth factor (VEGF); Protein A or D; Rheumatic factor; Bone-derived neurotrophic factor (BDNF); Neurotrophin-3, -4, -5, or -6 (NT-3, NT-4, NT-5, or NT-6); NGF-beta; Platelet-induced growth factor (PDGF); aFGF; bFGF; Epithelial cell growth factor (EGF); TGF-alpha; TGF-beta, such as TGF-beta1, TGF-beta2, TGF-beta3, TGF-beta14, or TGF-beta5; Insulin-like growth factor-I or -II (IGF-I or IGF-II); des(l-3)-IGF-I (brain IGF-I), insulin-like growth factor binding protein, erythropoietin; Bone inducing factor; Immunotoxin; Bone morphogenetic protein (BMP); Interferon-alpha, -beta, or -gamma; M-CSF, GM-CSF or G-CSF; IL-1 to IL-10; Superoxide dismutase; Decay accelerating factor; AIDS envelope protein; Transport protein; Homing receptor; Addressin; Regulatory protein; CD3, CD4, CD8, CD11a, CD11b, CD11c, CD18, CD19, CD20, CD34, CD40, or CD46, ICAM, VLA-4 or VCAM; Or a HER2, HER3 or HER4 receptor; A constituent of the ErbB receptor family; EGF receptor; HER2, HER3 or HER4 receptor; Cell adhesion molecule; LFA-1, Mac1, p150.95, VLA-4, ICAM-1, VCAM, alpha4/beta7 intigrin or alphav/beta3 intigrin; Alpha or beta subunits of cell adhesion molecules; Antibodies); Growth factor, VEGF; Tissue factor (TF); TGF-beta; Alpha interferon (alpha-IFN); IL-8; IgE; Blood type antigen Apo2, death receptor; flk2/flt3 receptor; Obesity (OB) receptor; mpl receptor; CTLA4 or protein C.

In another embodiment, the present invention provides antigen binding fragments obtainable by the methods described herein. This antigen-binding fragment can be used, for example, for therapeutic or diagnostic purposes.

The sequences used in the following examples are as follows:

> 903 or 578minmax (SEQ ID NO: 1)

> 791 or 34max (SEQ ID NO: 2)

> scFvl 05 (SEQ ID NO: 3)

> 961 or 578minmaxDHP (SEQ ID NO: 4)

Example 1

Antidrug antibody bridge ELISA (ADA-ELISA)

1.1. background

Antibodies pre-existing for monoclonal antibodies may be against constant regions, variable domain framework positions or antigen binding loops, CDRs. Pre-existing antibodies that specifically bind to Fv fragments rather than IgG have the potential to recognize regions previously shielded from IgG. These regions are primarily the domain interface located between the variable domain and constant region 1 (CL1 or CH1) or the Fab portion and the Fc domain (CH2 and CH3). Antibodies that recognize these interfaces are in all possibilities a specific format. Because the framework sequence of scFv903 is highly conserved in humans, pre-existing antibodies to scFv903 in human serum are likely to bind to CDR or V/C-boundary residues. The epitope of these pre-existing anti-scFv903 antibodies was characterized by evaluating the potential of various scFvs competing with the binding of anti-drug antibody (ADA) and ESBA903 in a sandwich ELISA. The tested scFvs are:

-scFv containing the same framework as scFv903 but different CDRs (34_max(791)),

-scFv having a framework different from scFv903 and a different CDR (scFvl05), and

-ScFv903 variant containing a substitution at the previous V/C interface (scFv903 DHP (961))

The ELISA developed for the screening of anti-scFv903 antibodies is a semi-quantitative assay and was developed in a bridge format capable of detecting responses of all antibody isotypes from different species (see Fig. 1).

Briefly with reference to FIG. 1, microtiter plates were coated with scFv 903 1, 2 to which samples containing anti-scFv 903 antibody 3,6 were bound. As a first detection reagent, a biotinylated scFv 903 4 was used to detect the conjugated scFv 903/antiscFv 903 complex, followed by detection with a second detection reagent 4 and streptavidin Poly-HRP 5. The amount of anti-scFv 903 antibody present in the quality control and sample was measured using a peroxidase (POD) substrate (3,3'-5,5' tetramethylbenzidine (TMB)).

Development of ADA ELISA was performed with the positive control antibody AB903-3. Rabbits were immunized with scFv 903, followed by affinity purification of serum (Squarix Biotechnology) to develop an anti-scFv 903 antibody stock (rabbit polyclonal anti-scFv 903 IgG called AB903-3). As shown in Fig. 1B, the epitope of the pre-existing antibody on scFv 903 was characterized by competition between ADA bound to scFv 903 and the scFv described above.

1.2 Essay method

Microtiter plates (Nunc Maxisorp) were coated with 0.1 mcl/ml scFv 903 in PBS (Dulbecco, Sigma). The sealed plate was incubated overnight at 4°C.

The plate was washed 3 times with 300 mcl per well of washing buffer (TBST 0.005% Tween (20)/Atlantis Microplate Washer (ASYS)). Non-specific sites were blocked with 280 mcl per well of blocking buffer (PBS, 10 mg/ml BSA 1% (w/v), 0.1 ml/50 ml Tween 20 (0.2%, v/v)) Sealed plate at room temperature (25) C) with shaking for 1.5 hours, then the plate was washed again 3 times as indicated above.

An analyte control (affinity purified rabbit polyclonal anti-scFv 903 IgG or human serum called AB903-3) was added at three different concentrations:

HiQC: 2500 ng/ml AB903

MeQC: 500 ng/ml AB903

LoQC: 250 ng/ml AB903

QC was mixed into each NSB serum collection (all serum collections used for determination of the assay cut point, >30). The sample to be measured was diluted in a ratio of 1 to 10. A sample of 50 mcl per well was added dropwise and the sealed plate was incubated at room temperature (25° C.) for 2.0 hours.

As described above, the plate was washed 3 times with wash buffer. As a first detection reagent, biotinylated scFv 903 (500 mcg protein biotinylated with Lightning-Link kit (protocol: Lightning-Link™ Biotin Conjugation Kit, Type A, #704-0015, Innova Biosciences)) was added. For the above purposes, biotinylated scFv 903 was added to a dilution buffer at a concentration of 250 ng/ml (PBS, 10 mg/mL BSA 1% (w/v), 0.1 ml/50 ml Tween 20 (0.2% v/v)). Diluted with. 50 mcl per well was added. The sealed plate was incubated at room temperature (approximately 25° C.) for 1.0 hour with shaking.

The plate was washed again 3 times as described above. A second detection reagent, streptavidin-Poly-HRP (Stereospecific Detection Technologies, 1 mg/ml) was diluted 1:5'000 with a dilution buffer, and 50 mcl per well was added. The sealed plate was incubated at room temperature (approximately 25° C.) for 1.0 hour with shaking.

For detection, the plate was washed 3 times with 300 mcl washing buffer per well as described above. The plate was then washed twice with 300 mcl of ddH20 per well. Thereafter, 50 mcl POD (TMB) substrate per well was added at room temperature. After incubation for 3-6 minutes (incubation time should not exceed 30 minutes at most), 50 mcl/well 1M HCL was added to stop the reaction. If the color development reaction is too strong, the reaction is stopped by adding 50 mcl/well 1M HCL sooner.

Plates were read at 450 nm using a Tecan Sunrise microtiter plate reader. Typically, the reaction was repeated 3 times for each quality control, NSB and individual serum sample. The readings were averaged.

1.3 Essay Cut Point (ACP) Determination

Essay cut points were established during the essay run for determination of positive. The cut point of the assay is the level of the assay response, above it the sample is defined as positive and below it the sample is defined as negative. Using a risk-based approach, a 5% false positive was more appropriate than a false negative. This was done using the average absorbance plus the standard deviation of 1.645 as a parametric approach, where 1.645 is the 95th percentile of the normal distribution. All individuals with OD ≥ 3* standard deviation were excluded and a new essay cut point was calculated. If the OD of up to 5% of the individuals is above the ACP, the calculated value can be used as the ACP. On the other hand, the process was repeated until the maximum 5% of individuals had an OD of ACP or higher by reapplying the exclusion criteria for the remaining individuals.

After excluding 34 of the 149 individual sera, the assay cut point was corrected to 0.110. 40 serums (26.8%) showed a higher OD than the assay cut point (see Fig. 2). LoQC, MeQC and HiQC samples were included in the monitoring of assay performance. For the preparation of NSB, 34 sera transferred for statistical evaluation of assay cut points were excluded from the serum collection. The average NSB result was 0.073, and the normalization factor was 1.50.

1.4 Determination of normalization factor and plate specific cut point

The ACP was determined as NSB, then a normalization factor was applied to calculate the plate specific cut point for subsequent measurements with the same NSB. The OD value of NSB was evaluated to determine the normalization factor. Three replicates of NSB (triplicate) were analyzed on each plate. The normalization factor is defined as the assay cut point divided by the average absorbance of the NSB. Plate specific cut points for each plate were calculated as follows:

Plate specific cut point = NSB absorbance * normalization factor

1.5 Confirmation Essay

In the case of detected ADA in serum, the confirmation assay demonstrates that the antibody found to be positive in the ADA bridging ELISA is specific for scFv 903. The validation assay is similar to the screening assay except that positive samples are mixed and pre-incubated with assay buffer containing scFv 903 or only assay buffer prior to analysis. For this purpose, the control material was mixed in dilution buffer or human serum at concentrations of LoQC, MeQC and HiQC levels. The samples were then diluted 1/2 with buffer or buffer containing scFv 903 (for human serum) or 10 mcg/ml, 100 mcg/ml or 1 mcg/ml. Samples were incubated for about 60 minutes at RT to bind scFv 903 to the ADA present in the samples. Samples were further diluted in buffer before loading onto the plate to minimize total matrix dilution. scFv 903 interferes with the binding of ADA to scFv 903 coated on the plate (see Figure 1). Therefore, a change in OD value of> 30% between serum diluted 1/2 with buffer and serum diluted 1/2 with buffer containing scFv 903 is minimal inhibition confirming the presence of specific anti-scFv 903 antibody. Is defined as Pre-incubation with scFv 903 at 10 mcg/ml, 100 mcg/ml as well as 1 mcg/ml resulted in an OD change of 60% to 95% for all 3 QC levels tested. A concentration of 100 mcg/ml was chosen for the confirmation assay.

1.6 Competition Essay Results

In order to map the binding site of the anti-scFv 903 antibody, the IgG format of scFv 903 as well as a set of different scFvs having a known amino acid sequence were used in the above identification assay setup instead of an excess of scFv 903. If in this experimental setup the ADA also recognizes a similar epitope on the test scFv, a given test antibody can only compete for binding of scFv 903 to the anti-scFv 903 antibody (ADA). Thus, signal reduction in the assay indicates the presence of at least one epitope shared between scFv 903 and test scFv. The following test antibodies were used in the experiments: scFv105, a humanized TNF-inhibiting scFv antibody fragment comprising mouse CDRs grafted onto a human scFv scF of type Vk1-VH1b; a humanized TNF-inhibitory antibody fragment comprising rabbit CDRs grafted to the same scFv scaffold as used in scFv 791, scFv 903 (Vkl-VH3); scFv 961, a derivative of scFv 903 comprising three point mutations (DHP motif) in the region located at the interface between the variable and constant domains, and the IgG format of scFv903-IgG, scFv 903.

The correlation of sequence variations between the four test molecules with differences in ADA binding properties was used to identify the entire scFv scaffold and, more specifically, the ADA epitope at the V-C interface. In addition, the format specificity of the pre-existing ADA was further confirmed by using competition with the full-size version of scFv903 (IgG).

Data from competition experiments are summarized in Table 1. The binding of all except the two individual human serums was competed with an excess of scFv 903 to confirm that this ADA was specific for ESBA903. Of the 32 human serums specific for scFv 903, only two are not competed by scFv791, so the antibodies present in these human serums (H53 and H76) bind to the scFv903 CDRs, but all other sera are clearly not CDR specific. Shown. Although this response is slightly lower, about 48% of ADA also recognizes the epitope on scFv105. This indicates that most of the antibodies are not explicitly framework specific, but rather bind conserved amino acids in different scFv scaffolds. Interestingly, the IgG format of scFv 903 did not seriously compete with scFv 903 in binding to the serum tested. This strongly suggests that most of the serum binds to the interface between the variable and constant regions and thus has access to the ADA of scFv 903 and not in the IgG format of scFv 903. In addition, scFV903 and scFv 961, which differed only by 3 amino acids, competed for binding of only 64% ADA. This reaction is generally lower than that of scFV903, indicating that the major fraction of pre-existing ADA binds to an epitope comprising the three amino acids that make up the hydrophobic surface patch at the V-C interface of scFv 903. The difference between the results obtained with scFV105 compared to scFV903 can be explained by the presence of different patterns of hydrophobic surface patches in these two molecules. In summary, these results indicate that up to 50% of pre-existing ADA in human serum are scFv format specific antibodies.

Table 1:

Epitope characterization of pre-existing antibodies in human serum. 100 mcg/ml of scFv903, 791 (34_max), scFv105 and 961 (scFv903_DHP) and percent reduction in OD when competing with the IgG format of scFv903 is provided for 34 different human sera with pre-existing anti-scFv903 antibodies.

To identify the site of interaction between ADA and scFv 903, sequence variations between scFv 903 and other scFvs (791, 105 and 961) were correlated with the specificity of ADA in various human serums. In the first step, the sequences of different scFvs were aligned and the solvent exposure sites were grouped according to the sequence difference between scFv 903 and other scFvs (FIGS. 3 and 4 ). Here, "α" represents a group of different positions between scFv 961 and all other scFvs, "β" represents a position different from that of the sequence from other molecules only scFv 105, and "γ" is all other scFv 791 And indicate a different position. In addition, αβ and αγ represent the conserved positions in all scFvs except for scFv 961 and scFvl05 or scFv961 and scFv791 respectively. Similarly, human sera containing anti-scFv903 antibodies were sorted by their specificity for other scFvs measured in competition assays using the same classification code used above for amino acid positions (see Table 2). A 50% minimum signal reduction in the competition assay was set as the threshold to fit the serum into binding to a given test scFv. "Α" in this test represents human anti-scFv 903 serum that did not show binding activity to scFv 961. “β”-serum did not bind scFv105, and “γ”-serum did not bind scFv 791. From the correlation between sequence analysis and in vitro binding tests, it can be concluded that, for example, an anti-scFv 903 antibody of type "α" human serum interacts with at least one amino acid from the amino acid group "α". Likewise, different types of serum interact with at least one amino acid within each amino acid group.

Structural analysis and homology model of scFv 903 were performed using Discover Studio version 2.5.5. The modeled structure was analyzed to determine which amino acid residues were exposed to the solvent and which amino acid residues were buried. The maximum possible solvent of each residue was calculated by measuring the relative Solvent Accessible Surface (SAS) of each residue in relation to the accessible surface area. Since the cutoff was defined as 25%, residues with a relative SAS of 25% or more were considered solvent exposed.

94% of the serum had antibodies specifically bound to scFv 903 or scFv 791, indicating that most of the pre-existing antibodies did not bind to the CDR regions. Half (50%) of human serum showed little or no antibody binding to scFv 961. Sequence analysis showed that scFv 961 and scFv 903 differ only in positions 12, 103 and 144 of the variable heavy chain. Thus, L12, V103 and L144 in the variable heavy chain of scFv 903 are involved in ADA binding (Table 2 and FIGS. 4 and 5). This finding was also confirmed by the fact that the IgG format of scFv 903, which is solvent inaccessible due to contact of each interface residue with neighboring constant regions, does not compete with the binding of scFv 903 with anti-scFv 903 serum.

12% of the tested human serum had antibodies against all antigenic regions (α,β,γ).

64% of serum contained little or no antibody that binds to scFv 961 (α). As mentioned above, this scFv differs from scFv 903 only at 3 residue positions (L12S, V103T, L103T) located at the previous variable-constant domain interface. Since these residue positions at the V/C interface are highly conserved, mutating these positions already provides a general way to deimmunize scFvs without affecting their biophysical properties.

51% of the serum contained little or no antibodies to scFv of the scFvl05 (β) Vk1-VH1b subtype (different framework). Possible antigenic regions were identified by sequence alignment (Figs. 3, 4 and 5).

23% of serum was specific for scFv 903, but did not bind scFv105 or scFv 961 (αβ). There is only one residue position that is different in the two scFvs compared to scFv903. Thus, each amino acid in the variable heavy domain of scF 903 (L12) plays an important role in the interaction between ADA and scFv 903.

In summary, a) approximately 50% of the ADA binds to the epitope at the interface between the variable region and the constant region (α), and contains residues at positions 12, 103 and 144 in the variable heavy domain, and b) these three highly It can be concluded that mutating conserved residues significantly reduces the binding strength and, in general, the frequency of pre-existing antibodies to scFv. This comprehensive applicability of the DHP-motif to reduce the immunogenicity of scFv is also supported by the high frequency of leucine at position 12, valine at position 103 and leucine at position 144 in variable heavy chains of human origin (FIG. 6 ). Because the binding of anti-scFv 903 serum to scFv105 is generally weaker compared to scFv 903, substitution of solvent-exposed residues within scFv 903 for each amino acid in scFv105 can potentially eliminate or attenuate the B-cell epitope. In particular, residue numbers 101 and 148 of the variable light chain are noted, because these bulky residues participate in the previous constant-variable domain interface (Table 4).

Although preferred embodiments of the present invention have been shown and described so far, it will be apparent that the present invention is not limited thereto, and on the other hand, can be variously implemented and practiced within the scope of the following claims.

SEQUENCE LISTING <110> ESBATech an Alcon Biomedical Research Unit Urech, David Gunde, Tea Borras, Leonardo <120> METHOD FOR DECREASING IMMUNOGENICITY <130> 3763 <160> 4 <170> PatentIn version 3.4 <210> 1 <211> 251 <212> PRT <213> Artificial <220> <223> antibody fragment <400> 1 Glu Ile Val Met Thr Gln Ser Pro Ser Thr Leu Ser Ala Ser Val Gly 1 5 10 15 Asp Arg Val Ile Ile Thr Cys Gln Ala Ser Glu Ile Ile His Ser Trp 20 25 30 Leu Ala Trp Tyr Gln Gln Lys Pro Gly Lys Ala Pro Lys Leu Leu Ile 35 40 45 Tyr Leu Ala Ser Thr Leu Ala Ser Gly Val Pro Ser Arg Phe Ser Gly 50 55 60 Ser Gly Ser Gly Ala Glu Phe Thr Leu Thr Ile Ser Ser Leu Gln Pro 65 70 75 80 Asp Asp Phe Ala Thr Tyr Tyr Cys Gln Asn Val Tyr Leu Ala Ser Thr 85 90 95 Asn Gly Ala Asn Phe Gly Gln Gly Thr Lys Leu Thr Val Leu Gly Gly 100 105 110 Gly Gly Gly Ser Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser Gly Gly 115 120 125 Gly Gly Ser Glu Val Gln Leu Val Glu Ser Gly Gly Gly Leu Val Gln 130 135 140 Pro Gly Gly Ser Leu Arg Leu Ser Cys Thr Ala Ser Gly Phe Ser Leu 145 150 155 160 Thr Asp Tyr Tyr Tyr Met Thr Trp Val Arg Gln Ala Pro Gly Lys Gly 165 170 175 Leu Glu Trp Val Gly Phe Ile Asp Pro Asp Asp Asp Pro Tyr Tyr Ala 180 185 190 Thr Trp Ala Lys Gly Arg Phe Thr Ile Ser Arg Asp Thr Ser Lys Asn 195 200 205 Thr Val Tyr Leu Gln Met Asn Ser Leu Arg Ala Glu Asp Thr Ala Val 210 215 220 Tyr Tyr Cys Ala Gly Gly Asp His Asn Ser Gly Trp Gly Leu Asp Ile 225 230 235 240 Trp Gly Gln Gly Thr Leu Val Thr Val Ser Ser 245 250 <210> 2 <211> 254 <212> PRT <213> Artificial <220> <223> antibody fragment <400> 2 Met Glu Ile Val Met Thr Gln Ser Pro Ser Thr Leu Ser Ala Ser Leu 1 5 10 15 Gly Asp Arg Val Ile Ile Thr Cys Gln Ser Ser Gln Ser Val Tyr Gly 20 25 30 Asn Ile Trp Met Ala Trp Tyr Gln Gln Lys Ser Gly Lys Ala Pro Lys 35 40 45 Leu Leu Ile Tyr Gln Ala Ser Lys Leu Ala Ser Gly Val Pro Ser Arg 50 55 60 Phe Ser Gly Ser Gly Ser Gly Ala Glu Phe Ser Leu Thr Ile Ser Ser 65 70 75 80 Leu Gln Pro Asp Asp Phe Ala Thr Tyr Tyr Cys Gln Gly Asn Phe Asn 85 90 95 Thr Gly Asp Arg Tyr Ala Phe Gly Gln Gly Thr Lys Leu Thr Val Leu 100 105 110 Gly Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser 115 120 125 Gly Gly Gly Gly Ser Glu Val Gln Leu Val Glu Ser Gly Gly Gly Leu 130 135 140 Val Gln Pro Gly Gly Ser Leu Arg Leu Ser Cys Thr Ala Ser Gly Phe 145 150 155 160 Thr Ile Ser Arg Ser Tyr Trp Ile Cys Trp Val Arg Gln Ala Pro Gly 165 170 175 Lys Gly Leu Glu Trp Val Ala Cys Ile Tyr Gly Asp Asn Asp Ile Thr 180 185 190 Pro Leu Tyr Ala Asn Trp Ala Lys Gly Arg Phe Pro Val Ser Thr Asp 195 200 205 Thr Ser Lys Asn Thr Val Tyr Leu Gln Met Asn Ser Leu Arg Ala Glu 210 215 220 Asp Thr Ala Val Tyr Tyr Cys Ala Arg Leu Gly Tyr Ala Asp Tyr Ala 225 230 235 240 Tyr Asp Leu Trp Gly Gln Gly Thr Leu Val Thr Val Ser Ser 245 250 <210> 3 <211> 245 <212> PRT <213> Artificial <220> <223> antibody fragment <400> 3 Asp Ile Val Met Thr Gln Ser Pro Ser Ser Leu Ser Ala Ser Val Gly 1 5 10 15 Asp Arg Val Thr Leu Thr Cys Thr Ala Ser Gln Ser Val Ser Asn Asp 20 25 30 Val Val Trp Tyr Gln Gln Arg Pro Gly Lys Ala Pro Lys Leu Leu Ile 35 40 45 Tyr Ser Ala Phe Asn Arg Tyr Thr Gly Val Pro Ser Arg Phe Ser Gly 50 55 60 Arg Gly Tyr Gly Thr Asp Phe Thr Leu Thr Ile Ser Ser Leu Gln Pro 65 70 75 80 Glu Asp Val Ala Val Tyr Tyr Cys Gln Gln Asp Tyr Asn Ser Pro Arg 85 90 95 Thr Phe Gly Gln Gly Thr Lys Leu Glu Val Lys Arg Gly Gly Gly Gly 100 105 110 Ser Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser Ser Gly Gly Gly Ser 115 120 125 Gln Val Gln Leu Val Gln Ser Gly Ala Glu Val Lys Lys Pro Gly Ala 130 135 140 Ser Val Lys Val Ser Cys Thr Ala Ser Gly Tyr Thr Phe Thr His Tyr 145 150 155 160 Gly Met Asn Trp Val Arg Gln Ala Pro Gly Lys Gly Leu Glu Trp Met 165 170 175 Gly Trp Ile Asn Thr Tyr Thr Gly Glu Pro Thr Tyr Ala Asp Lys Phe 180 185 190 Lys Asp Arg Phe Thr Phe Ser Leu Glu Thr Ser Ala Ser Thr Val Tyr 195 200 205 Met Glu Leu Thr Ser Leu Thr Ser Asp Asp Thr Ala Val Tyr Tyr Cys 210 215 220 Ala Arg Glu Arg Gly Asp Ala Met Asp Tyr Trp Gly Gln Gly Thr Leu 225 230 235 240 Val Thr Val Ser Ser 245 <210> 4 <211> 251 <212> PRT <213> Artificial <220> <223> antibody fragment <400> 4 Glu Ile Val Met Thr Gln Ser Pro Ser Thr Leu Ser Ala Ser Val Gly 1 5 10 15 Asp Arg Val Ile Ile Thr Cys Gln Ala Ser Glu Ile Ile His Ser Trp 20 25 30 Leu Ala Trp Tyr Gln Gln Lys Pro Gly Lys Ala Pro Lys Leu Leu Ile 35 40 45 Tyr Leu Ala Ser Thr Leu Ala Ser Gly Val Pro Ser Arg Phe Ser Gly 50 55 60 Ser Gly Ser Gly Ala Glu Phe Thr Leu Thr Ile Ser Ser Leu Gln Pro 65 70 75 80 Asp Asp Phe Ala Thr Tyr Tyr Cys Gln Asn Val Tyr Leu Ala Ser Thr 85 90 95 Asn Gly Ala Asn Phe Gly Gln Gly Thr Lys Leu Thr Val Leu Gly Gly 100 105 110 Gly Gly Gly Ser Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser Gly Gly 115 120 125 Gly Gly Ser Glu Val Gln Leu Val Glu Ser Gly Gly Gly Ser Val Gln 130 135 140 Pro Gly Gly Ser Leu Arg Leu Ser Cys Thr Ala Ser Gly Phe Ser Leu 145 150 155 160 Thr Asp Tyr Tyr Tyr Met Thr Trp Val Arg Gln Ala Pro Gly Lys Gly 165 170 175 Leu Glu Trp Val Gly Phe Ile Asp Pro Asp Asp Asp Pro Tyr Tyr Ala 180 185 190 Thr Trp Ala Lys Gly Arg Phe Thr Ile Ser Arg Asp Thr Ser Lys Asn 195 200 205 Thr Val Tyr Leu Gln Met Asn Ser Leu Arg Ala Glu Asp Thr Ala Thr 210 215 220 Tyr Tyr Cys Ala Gly Gly Asp His Asn Ser Gly Trp Gly Leu Asp Ile 225 230 235 240 Trp Gly Gln Gly Thr Thr Val Thr Val Ser Ser 245 250

Claims (10)

The variable heavy chain amino acid position 103 (AHo numbering) is substituted with serine (S) or threonine (T) with variable light chain amino acid position 99 (AHo numbering) to serine (S), variable heavy chain amino acid position 12 (AHo numbering) ) And substituting variable heavy chain amino acid positions 144 (AHo numbering) with serine (S) or threonine (T). 2. The method of claim 1, wherein the variable domains are part of a single chain antibody (scFv). The method of claim 1 wherein the amino acid residue to be substituted is leucine (L), valine (V), aspartic acid (D), phenylalanine (F), arginine (R) and / or glutamic acid (E). 4. The method according to any one of claims 1 to 3, wherein the variable heavy chain amino acid residue is
(a) serine (S) at heavy chain amino acid position 12;
(b) threonine (T) at heavy chain amino acid position 103; And
(c) substituted at the heavy chain amino acid position 144 with threonine (T). An antigen binding fragment obtained by the method according to claim 1. 6. The antigen-binding fragment of claim 5 for use in therapeutic or diagnostic use. 7. An antigen-binding fragment according to claim 6, which is applied locally and / or locally. 7. An antigen-binding fragment according to claim 6, which is applied as a sustained release preparation and / or device. 9. An antigen-binding fragment according to any one of claims 5 to 8, which is a single chain antibody (scFv). A variable light chain comprising serine (S) at light chain amino acid position 99 (AHo numbering); And
Serine (S) at heavy chain amino acid position 12 (AHo numbering), serine (S) or threonine (T) at heavy chain amino acid position 103 (AHo numbering) and serine (S) or threonine at heavy chain amino acid position 144 T);
(ScFv). &Lt; / RTI &gt;

Images